Recent animal and human neuroimaging studies suggest that antipsychotic medications are potentially toxic to cortical brain structures. However, the human studies have not been placebo-controlled and cannot establish causality. Also, some studies suggest that antipsychotics are actually protective against structural brain changes associated with untreated psychotic disorders. With the growing number of younger and older patients receiving antipsychotics, determining the overall effect of these medications on brain structure is a public health imperative. For ethical reasons, this question cannot be resolved in patients with schizophrenia. We are proposing to conduct a pre/post MRI/DTI neuroimaging study in the context of STOP- PD II, a newly NIMH-funded randomized, placebo-controlled trial of olanzapine in the continuation/maintenance treatment of major depression with psychotic features (psychotic depression, PD). The main aims of this application are to assess brain changes in gray and white matter associated with olanzapine treatment. Using a mixed model regression analysis, we will compare changes in brain structure and connectivity over time in patients in whom olanzapine has been continued or discontinued. We hypothesize that, compared to placebo, continuing treatment with olanzapine will be associated with: cortical thinning throughout all lobes; increase in striatal volumes; reductions in white matter microstructure; and no change in surface area or hippocampal and amygdala volumes. Our application leverages the clinical infrastructure (i.e., screening, recruitment, treatment, and clinical monitoring) provided by STOP-PD II: 152 subjects who achieve remission after 12 weeks of acute open treatment with sertraline plus olanzapine and remain stable over the following 8 weeks will be randomized to continuing olanzapine vs. being switched to placebo. These subjects will undergo a first (pre) MRI scan and will be rescanned (post) at the completion of the 36-week double-blind trial or at the time of relapse. This study is innovative and unique in its use of a randomized placebo-controlled design combined with advanced neuroimaging techniques: high resolution MRI (including DTI), measurements of cortical thickness, cortical surface area, microstructural integrity of white matter tracts, and subcortical striatal and limbic morphometry. It will extend and advance conventional volumetric approaches that have been used in humans and animals to assess longitudinal effects of antipsychotics on brain structure, providing a more biologically meaningful assessment of both gray and white matter.